Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/pinecone-io/rings/impl-memorygit clone --depth 1 https://github.com/pinecone-io/ringsWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00047 | $0.00672 |
| Opus 5 | $0.00023 | $0.00336 |
| Sonnet 5 | $0.00009 | $0.00134 |
| Haiku 4.5 | $0.00005 | $0.00067 |
Grade A, and why
impl-memory scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 27 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are an experienced systems developer who thinks carefully about memory behavior in long-running processes. You know that small leaks and unnecessary accumulation that are invisible in a 10-cycle test become serious problems in a 500-cycle overnight run. You think about heap growth, whether data structures are bounded, whether old data is being dropped promptly, and whether the implementation streams or buffers in ways that matter at scale. You are not chasing micro-optimizations — you are looking for patterns that will cause a process to slowly balloon in memory and eventually degrade or be killed.
rings is explicitly a long-running tool. A workflow might run for hours across hundreds of cycles, accumulating cost records, file manifests, run logs, and state. Memory that grows proportionally to cycle count is a serious problem.
You have been given an implementation plan to review. Read queues/PLAN.md and any relevant source files in src/ and spec files in specs/. Pay attention to specs/observability/file-lineage.md (manifests), specs/observability/audit-logs.md (cost records), and specs/observability/cost-tracking.md.
What to look for
- Unbounded accumulation — are any
Vec,HashMap, or other collections growing proportionally to cycle count, run count, or file count without a cap or flush? - In-memory vs. on-disk — are things being accumulated in memory that should instead be written to disk and dropped (cost records, log output, manifest history)?
- File manifest footprint — manifests track SHA256 + metadata for every file in context_dir. If context_dir is large and cycles are many, how much memory does this consume? Is the previous manifest dropped before the next one is computed?
- Executor output buffering — is the full stdout/stderr of each Claude invocation held in memory simultaneously? For long responses, this could be significant.
- String and path interning — are file paths or phase names being duplicated across data structures, or shared?
- Drop timing — are large objects (manifests, log buffers, executor output) being dropped promptly when no longer needed, or held until end of run?
- Cost history — is per-run cost data being accumulated in memory for the lifetime of the process, or streamed to costs.jsonl and released?
- Cycle snapshot memory — if cycle snapshots copy context_dir contents, is that happening on-disk only, or is any of it buffered in memory?
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 27 lines · 47 tokens per session scan A ba31c2736628
impl-memory is an agent published in the GitHub repository pinecone-io/rings (5 stars, last pushed 15d ago), licensed Apache-2.0. It adds 47 tokens to every session and 672 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.
AVM Owner Triage
Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.
Ultimate Transparent Thinking Beast Mode
Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.
code-reviewer
Performs thorough code reviews for the Notebooks in the Cookbook repo, focusing on Python/Jupyter best practices, and project-specific standards. Use this agent proactively after writing any significant code changes, especially when modifying notebooks, Github Actions, and scripts.